通信受限下网络化多传感器融合系统的分布式估计

With the rapid development of network communication technology, networked multi-sensor fusion systems (NMFSs) find wide application in military and civilian areas because of their advantages such as flexible architectures, simpler installation, easier maintenance and low cost. Therefore, NMFSs have now been one of the hot research topics in system and control area. However, the conventional fusion estimation approaches are not be applicable to the NMFSs, thus an important and practical problem is how to design fusion estimation algorithm for the NMFSs. This project is concerned with the distributed estimation problem for the NMFSs with communication constraints from two aspects in communication networks: one is the information transmission characteristics; the other is the form of information representation. The main study is summarized as follows: 1) propose the distributed estimation theory and algorithm for the NMFSs with Markov delays and packet dropouts; 2) propose the distributed estimation theory and algorithm for the NMFSs under the dual data compression strategy; 3) propose distributed estimation theory and algorithm for the NMFSs under the minimum traffic. Based on the designed distributed fusion estimators, the relationship between the parameters on the delay- packet-dropout and estimation performance is established, while the relationship between the dual data compression operator and estimation performance is also established. Moreover, the minimum traffic and the optimal bandwidth allocation are derived such that the designed estimator meets the satisfactory estimation performance. The proposed project will achieve some novel research results with respect to the distributed estimation theory in NMFSs, and thus promote the developments and applications of the conventional multi-sensor fusion estimation theory.

随着网络通信技术的飞速发展，网络化多传感器融合系统(简称NMFSs)以其布线少、成本低、易于扩展和维护等优点，在军事和民用领域有着重要的应用价值，目前是一个热点研究课题。然而，通信受限使得传统的融合估计理论无法直接应用于NMFSs，迫切需要提出适用于NMFSs的估计理论。本项目从网络环境中信息传输特性和信息表示模式两个角度出发，提出通信受限下NMFSs的分布式估计理论与算法，主要研究：基于Markov通信时延与丢包的NMFSs分布式估计理论与算法；基于双重数据压缩策略的NMFSs分布式估计理论与算法；基于最小通信量的NMFSs分布式估计理论与算法。分别建立时延、丢包特性参数、双重数据压缩算子与所设计估计器性能之间的关系，并给出保证满意估计性能的最小通信量及最优带宽分配方案。本项目的研究将在通信受限下NMFSs的分布式估计理论上取得创新和突破，促进传统多传感器信息融合估计理论的发展和完善。

随着网络通信技术的飞速发展，网络化传感器融合系统以其布线少、成本低、易扩展和维护等优点，在军事和民用领域有着重要的应用价值。网络化传感器融合系统一般应用于监测实际系统的动态过程，通过传感器的量测数据，为用户提供有用的信息。由于多传感器融合估计方法可以提高系统的监测精度，因此网络化融合估计目前是一个热点研究课题。虽然网络化传感器融合系统拥有诸多优点，但是由于通信网络的引入，也带来了新的问题，且传统的融合估计方法不适于这类系统。特别地，在网络化多传感器融合系统中必须考虑由于通信网络导致的有限带宽、传输延迟与丢包等问题。因此，我们提出了一系列网络化融合估计算法以应对系统中出现的通信受限问题。研究结果为融合估计算法在网络化环境中的应用提供了理论保证，促进传统多传感器信息融合估计理论的发展和完善。